Chambers (2002) cites an example of the application of the apparent-time hypothesis. The study, carried out in central Canada, examined the sociolinguistic variable (wh), where the unvoiced labiovelar glide /hw/ loses phonemic status and merges with the corresponding voiced glide /w/. In this study, the oldest subjects seem to indicate a stable period for this variable, both the 70-79 year olds and those over 80 used the voiced variant where the unvoiced was "expected" 38.3 and 37.7% of the time, respectively. Each subsequent younger age cohort (10 years) shows a greater percentage of /w/ usage, with those 20-29 using /w/ 87.6% of the time and the teenagers using it 90.6% of the time. Notice that the deltas between the oldest two groups and between the youngest two groups are relatively small, 0.6% and 3.0%. Between these two extremes the rate of change between the groups is quite high, approximately 10% per age cohort. This pattern can be described as an initial stable period, followed by a period of rapid change, and a tailing off as the change nears completion. This S-curve pattern has been identified as characteristic for many types of linguistic changes.
Not all age-related variation is indicative of change in progress. It may be an age-graded variation. The applicability of the apparent-time hypothesis should be confirmed by real-time evidence, which actually samples the population over an extended period of time. This is the only true indicator of change in progress. Real-time evidence may come from a longitudinal study of a population or by replicating a study conducted at some relatively distant time and comparing the observations to those previously published.